To evaluate whether there is a relationship between steroid treatment and risk for osteonecrosis of the hip and knee in patients with severe acute respiratory syndrome (SARS). The hospital ethics committee approved the study, and all patients provided written informed consent. A total of 254 patients with confirmed SARS treated with steroids underwent evaluation with magnetic resonance (MR) imaging for osteonecrosis. Clinical profiles, joint symptoms, relevant past medical and drug history, steroid dose, and radiographic and MR imaging evidence of osteonecrosis and other bone abnormalities were evaluated. Mann-Whitney, Kruskal-Wallis, and Pearson exact chi(2) tests were performed, and univariate and multivariate logistic regression analyses were applied. One hundred thirty-four (53%) of 254 patients had recent onset of large joint pain, but 211 (80%) of 264 painful joints were not associated with abnormality on MR images. MR images in 12 (5%) of 254 patients showed evidence of subchondral osteonecrosis in the proximal femur (n = 9), distal femur (n = 2), and proximal and distal femora and proximal tibiae (n = 1). Additional nonspecific subchondral and intramedullary bone marrow abnormalities were present in 77 (30%) of 254 patients. Results of multiple logistic regression analysis confirmed cumulative prednisolone-equivalent dose to be the most important risk factor for osteonecrosis. The risk of osteonecrosis was 0.6% for patients receiving less than 3 g and 13% for patients receiving more than 3 g prednisolone-equivalent dose. No relationship was found between additional nonspecific bone marrow abnormalities and steroid dose. An appreciable dose-related risk was found for osteonecrosis in patients receiving steroid therapy for SARS. Additional nonspecific bone marrow abnormalities were frequent. Joint pain was common after SARS infection and was not a useful clinical indicator of osteonecrosis. (c) RSNA, 2005.